How to Write a Nonprofit Website Redesign RFP (With a Template)

Most website redesign RFPs are wishlists. They describe the site the organization imagines, in its own vocabulary, and list features someone admired on a peer's site: a modern look, an events calendar, a bigger donate button, pages staff can edit without calling anyone. Then they ask vendors to price it.

One bidder assumes you will write the content and quotes for design and build. Another assumes it is writing the content and quotes several times as much. Neither mentions redirects, because the RFP never did, and nothing on either quote says which assumption produced which number. Whoever wins, the content is still unwritten in month three because nobody agreed who was writing it, and in month seven search traffic falls and nobody can find a list of the pages the old site had.

Any competent designer can answer "the site looks dated" and leave untouched the things that decide the outcome, such as whether the donation flow keeps people. Those are missing from the RFP because nobody measured them first.

Before anyone drafts the RFP, somebody spends a few weeks finding out what the current site does and writes it down in a form a stranger could check.

A wishlist says "we want a modern, mobile-friendly site with better donation conversion." An RFP built on a review says something like this: sixty-two percent of sessions are on a phone, the donation flow loses seven in ten people at the details step, eleven named URLs produce most of the search entrances and must keep working, and four forms send submissions to inboxes nobody reads. Every bidder is now pricing the same job.

Five measurements do most of the work. None needs a design opinion, and someone with admin access can pull each in an afternoon.

Then label every finding. Observed means somebody saw it happen or read it in a report. Inferred means a reasonable conclusion from what was seen. "The volunteer form sends to a shared inbox with nine hundred unread messages" is observed. "Volunteers drop out because nobody replies" is inferred. Both go in the RFP with their labels, so a bidder can price the certain work firmly and ask about the rest. The label also catches the organization's own unchecked beliefs, such as the certainty that most donors give online, before a vendor designs around them.

Three kinds of firm answer website redesign RFPs, and each is good at real work. A proposal is shaped by what its author considers the job, so the parts a firm sees as someone else's job are usually absent rather than declined.

A marketing firm that leaves the redirect map out is working to its scope, and most will say so if you ask. Trouble starts when the organization assumes one hire covers all three.

So give every left-out item its own line in the scope, even when the bidder will not own it, and beside it write "bidder" or the staff member or partner who will do it. A bidder who reads "redirect map: bidder" and prices it has told you something. So has one who skips the line.

If you are working out how to write a website RFP from nothing, start from this list, in roughly the order a bidder needs it.

Six terms belong in the RFP as requirements and in the contract in the same words.

The examples are for an invented organization, Anytown Literacy Project. Replace each with your own evidence. A row you cannot fill is where your review should start.

Send these to every bidder in writing and ask for written answers.

Lay the proposals side by side against the RFP's scope and ownership lines and mark each one included, excluded, or silent. Treat silent as excluded. A bid that says nothing about the redirect map does not include one, whatever is said on the call, and a bid far below the others has usually skipped a line this way.

Score the answers to the eight questions before anyone sees price. Two or three people score alone on a simple scale, then compare. Where scorers disagree there is usually an answer worth a follow-up call.

For references, ask for a site that launched at least eighteen months ago and ask about year two. Did traffic hold? Did the forms keep working? Could staff make changes without calling the firm? How did maintenance cost compare with the estimate?

With two finalists left, a paid discovery of a week or two shows how each works on your own problem. Put your review document at the center of it so neither starts from scratch at your expense. Then decide in the order you said you would, and tell the bidders you did not choose why.